EU Strikes AI Omnibus Deal: High-Risk Rules Delayed to December 2027
At 4:30 a.m. on Wednesday 7 May 2026, after negotiations that ran deep into the small hours, the Council of the European Union and the European Parliament reached a provisional political agreement on the AI Omnibus — the most consequential rewrite of EU artificial-intelligence rules since the original AI Act of 2024.
Although the final text has not yet been published, the agreement provides immediate clarity on a question that has dominated boardrooms across the continent for six months: when, exactly, will high-risk AI obligations actually start to bind? The answer, in the form most observers expected, is end of 2027 — a substantial delay from the 2 August 2026 deadline that would otherwise have applied.
Co-rapporteurs claim a balanced outcome
The co-rapporteurs in Parliament — Arba Kokalari (EPP, Sweden) and Michael McNamara (Renew, Ireland) — defended the package as a balanced response to two competing pressures: industry demands for legal certainty and timeline relief, and civil-society concerns about emerging AI harms.
“I’m pleased that this morning we reached an agreement on the AI Omnibus,” said McNamara at the press conference following the deal. “Alongside simplification measures, we are banning nudification apps, a key part of the Parliament’s mandate, and, of course, the creation of child sexual abuse material using AI systems. This way, we have the tools to act if providers do not address AI systems that compromise fundamental rights or human dignity.”
The European Commission, which proposed the Digital Omnibus on AI only five months earlier as part of the EU’s broader simplification agenda, welcomed the deal. In its press statement the Commission said the agreement “will make the implementation of the AI Act for EU businesses easier while maintaining its benefits for European society, safety and fundamental rights.”
What actually changes
Three substantive changes will matter most in practice. First, the application date for transparency and watermarking obligations — rules designed to make AI-generated audio, image, video and text detectable and traceable — has been moved from 2 August 2026 to 2 December 2026, with the grace period for providers reduced from six months to three.
Second, the start date for the obligations applicable to high-risk AI systems (Annex III use cases: biometrics, critical infrastructure, education, employment, migration, asylum, border control, administration of justice) is pushed to end of 2027, against fierce European Parliament resistance to any longer delay.
Third, and unprompted by the Commission’s original proposal, the package introduces a ban on “nudifier” apps — AI systems that generate non-consensual intimate imagery from clothed photographs. The Parliament’s mandate to add this prohibition was secured during the trilogue.
The Draghi context
The wider context for the Omnibus is the Draghi Report on EU competitiveness, which has driven Commission President Ursula von der Leyen’s second-mandate emphasis on regulatory simplification. Omnibus VII, of which the AI changes form part, is the seventh such simplification package since 2025 — and not the last.
Formal adoption is now expected before 2 August 2026, when the unmodified high-risk rules would otherwise have entered force. Upon adoption, the amendments will be published in the Official Journal of the European Union and enter into force three days later. For Europe’s AI industry, the next eighteen months will be defined less by the text of the law than by the speed and quality of the standards, codes of conduct and guidance that the Commission and the AI Office must now produce to make the revised regime operationally workable.
